【发布时间】:2015-02-11 21:40:47
【问题描述】:
我正在制作一个等值线图的小平面/点阵图,每个图都显示了不同的模型运行如何影响一个变量,该变量映射到多个多边形。问题是输出图形会在每个绘图的多边形之间产生奇怪的线条(见下图)。
虽然我已将 shapefile 操作并转换为具有 ggplot2 适当属性的数据框,但我不熟悉如何使用该包的详细信息,并且在线文档仅限于这样一个复杂的包。我不确定是什么参数导致了这个问题,但我怀疑它可能是 aes 参数。
脚本:
library(rgdal, tidyr, maptools, ggplot2, dplyr, reshape2)
setwd('D:/path/to/wd')
waterloo <- read.table("waterloo-data.txt", header=TRUE, sep=',', stringsAsFactors=FALSE)
waterloo <- data.frame(waterloo$DAUID, waterloo$LA0km, waterloo$LA4_exp, waterloo$LA20km, waterloo$LA30km, waterloo$LA40km, waterloo$LA50km)
colnames(waterloo) <- c("DAUID", "LA0km", "LA10km","LA20km", "LA30km", "LA40km", "LA50km")
## Produces expenditure measurements by ID variable DAUID, using reshape2/melt
wtidy <- melt(waterloo, id.vars=c("DAUID"), measure.vars = c("LA0km", "LA10km", "LA20km", "LA30km", "LA40km", "LA50km"))
colnames(wtidy) <- c("DAUID", "BufferSize", "Expenditure")
wtidy$DAUID <- as.factor(wtidy$DAUID) # for subsequent join with wtrl_f
### READ SPATIAL DATA ###
#wtrl <- readOGR(".", "Waterloo_DA_2011_new")
wtrl <- readShapeSpatial("Waterloo_DA_2011_new")
wtrl$id <- row.names(wtrl)
wtrl_f <- fortify(wtrl)
wtrl_f <- left_join(wtrl_f, wtrl@data, by="id")
# Join wtrl fortified (wtrl_f) to either twaterloo or wtidy
wtrl_f <- left_join(wtrl_f, wtidy, by="DAUID")
### PLOT SPATIAL DATA ###
ggplot(data = wtrl_f, # the input data
aes(x = long.x, y = lat.x, fill = Variable/1000, group = BufferSize)) + # define variables
geom_polygon() + # plot the DAs
geom_path(colour="black", lwd=0.05) + # polygon borders
coord_equal() + # fixed x and y scales
facet_wrap(~ BufferSize, ncol = 2) + # one plot per buffer size
scale_fill_gradient2(low = "green", mid = "grey", high = "red", # colors
midpoint = 10000, name = "Variable\n(thousands)") + # legend options
theme(axis.text = element_blank(), # change the theme options
axis.title = element_blank(), # remove axis titles
axis.ticks = element_blank()) # remove axis ticks
输出图形如下:
奇怪的!我取得了很好的进展,但我不知道 ggplot 从哪里得到这些线。对此的任何帮助将不胜感激!
PS;作为另一个不相关的问题,多边形线相当参差不齐。我将如何平滑这些线条?
【问题讨论】:
-
如果你能提供一个最小的、可重现的例子,帮助会容易得多。最小的意思是足够的代码和数据来显示问题。可重现的意思是我可以将您提供的代码完全按照提供的方式粘贴到新的 R 会话中并无错误地运行它(当然,您需要帮助的错误除外......)